openspec-plus-design

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill relies on external documentation files as inputs for its design logic, which could contain malicious instructions designed to influence the agent's output.
  • Ingestion points: Processes proposal.md, spec.md, AGENTS.md, CLAUDE.md, GEMINI.md, and configuration files like openspec/.plus/config.yaml found in SKILL.md.
  • Boundary markers: Does not implement specific boundary markers or instructions to ignore embedded prompts within the ingested files.
  • Capability inventory: The skill can execute shell commands via openspec, perform file writes to an outputPath, and dispatch subagents.
  • Sanitization: Content from external sources is used directly without defined sanitization or validation steps.
  • [COMMAND_EXECUTION]: The skill executes the openspec CLI tool (e.g., openspec instructions design --change <name> --json) to retrieve design templates and rules. The use of variables like <name> in the shell command represents a potential path for command argument injection if the calling agent does not correctly sanitize user-provided task names.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 09:06 AM
Security Audit — agent-trust-hub — openspec-plus-design